Apolipoprotein E (apoE) displays critical isoform-specific effects in neurodegeneration and in the normal maintenance and repair of neurons. Unlike the other major human isoforms, apoE3 and apoE2, apoE4 is an established risk factor for Alzheimer's disease (AD). However, the basis underlying this isoform-specific effect is unknown and, most importantly, has not been explored systematically in terms of the effect of structure on function. A basic paradigm of protein chemistry is that the structure and biophysical properties of a protein determine whether it functions normally or abnormally. Thus, analyzing the structural and biophysical differences among the isoforms can provide important clues regarding the apoE isoform-specific mechanisms and basis for the association of apoE4 with AD. Previous studies fi'om the Project Leader's laboratory identified three major characteristics that distinguish apoE4 from apoE3 and apoE2: (1) the amino-terminal domain of apoE4 is the least resistant to chemical or thermal unfolding and forms a stable folding intermediate, which we determined is a molten globule state; (2) apoE4 lacks cysteine and does not form a disulfide-linked homodimer, whereas apoE3 and apoE2 contain cysteine at position 112 and form dimers; and (3) apoE4 domain interaction, an interaction of the amino- and carboxyl-terminal domains that is unique to apoE4. Our central hypothesis is that one or more of these structural or biophysical differences plays a major role in the association of apoE4 with neurodegeneration or deficits in neuronal repair. Our experimental approach is to alter the mouse Apoe gene by gene targeting to "humanize" mouse apoE with respect to each of the human isoform structural differences by introducing mutations that engineer in these structural differences individually and selectively. Using mouse models expressing mutant apoE displaying selected structural and biophysical features of human apoE4, we will examine the relative contribution of each of the human isoform structural differences to apoE4 behavior. As proof of principle, we have generated a mouse model of apoE4 domain interaction by gene targeting and are characterizing its phenotype. In this proposal, we will extend this structure-based approach with three specific aims that will test the hypothesis that the propensity of apoE4 to form a molten globule state and its lack of cysteine also contribute to the apoE4-specific effects. The identification of the key apoE4 structural and biophysical differences responsible for neurodegeneration holds the potential to provide new opportunities for novel therapeutic strategies designed to interfere with or diminish the pathological impact of these differences.